What to write in referral letters from primary care physicians to psychiatrists

Contents
  1. Methods of the study
  2. Prioritize sufficient information over brevity, and include the patient’s state before onset and family relationships
  3. In the medication history, write the dose and duration as well as the drug name
  4. Write your own assessment, even if it is only a symptom name or a provisional diagnosis
  5. Write what you ask of the psychiatrist and what you will continue yourself
  6. The 18 items expected in referral letters
  7. Limitations of the study
  8. References

When a primary care physician refers a patient to a psychiatrist, the referral letter is the basic means of communication between the two. Even so, few studies have examined what should be written in it and in how much detail. One qualitative study asked both family physician specialists and psychiatrists what they think should be written in referral letters1. I took part in this study as a co-author of the paper.

In a clinic, a physician at a desk holds a pen and writes on paper. Two adults, a patient and a family member, sit side by side on chairs beside the desk. A pink envelope lies on the desk, with a window and a potted plant in the background
Writing a referral letter for a regular patient

Based on this study, my Take-home message consists of the following four points.

  1. Prioritize sufficient information over brevity, and include the patient’s state before onset and family relationships
  2. In the medication history, write the dose and duration as well as the drug name
  3. Write your own assessment, even if it is only a symptom name or a provisional diagnosis
  4. Write what you ask of the psychiatrist and what you will continue yourself
The referral loop and four points to write in the referral letter
Primary carePsychiatristReferral letter1State before onset and family2Drugs: name, dose, duration3Assessment, its basis and course4What I ask and what I continuePatientExplain to the patient why, and what to askReply and reverse referralAssessment and feedbackDiagnosis and treatment planReverse referral once stable

Methods of the study

The study was conducted from October 2024 to March 2025. The participants were 17 family physician specialists certified by the Japan Primary Care Association and 15 psychiatrists who work in an outpatient setting at least one day per week. Three groups of only family physician specialists and three groups of only psychiatrists were formed, with 4 to 6 people in each, and semi-structured focus group interviews were held online. Two questions were asked: what is written in their usual referral letters, and what an ideal referral letter would look like if there were sufficient time and optimal conditions. The transcripts were analyzed using framework analysis.

A laptop on a desk shows six faces arranged in two rows of three. Beside the laptop is a cup of coffee, and only the coffee is pink
Online discussions were held in groups separated by specialty

What was analyzed is what physicians said about what they write in referral letters and what information they want. The 18 items are a reference for choosing the necessary information according to the case and the purpose of the referral; the study did not test the effect of including them on care or outcomes. The quotes below are taken from the paper.

Prioritize sufficient information over brevity, and include the patient’s state before onset and family relationships

I think the starting point of a referral letter is “who has what problem.” In this study, primary care physicians said that they write specifically what the problem is and who has it, including whether it is the patient or the family or caregivers who are troubled.

Primary care physicians said that they try to include information gained through a continuing relationship, such as facts the patient or family have forgotten, the patient’s state before onset, and family relationships. Psychiatrists also saw value in these descriptions.

Ideally, I think it would be even more interesting if premorbid personality was included. The first time [psychiatrists] meet the patient is the time when he is sick, so we don’t know [how] the patient used to be. For example, if the patient were cheerful, bright, and full of jokes during a visit to the PCP for hypertension, it would be unusual if the patient were currently in a gloomy state. Because of things like that, I think it is [a] strength of PCP[s]. (Psychiatrist)

About the family, family relationships and dynamics came up in addition to family composition, the key person, and the family members who attend the consultation. Psychiatrists valued these descriptions because family relationships can be considered during consultations and because the information is required for involuntary hospitalization. The stated intention behind writing occupation, economic status, ADL, functioning, character, and interests was for the psychiatrist to tailor the treatment goal to the patient’s living circumstances. Physical information (past and present medical history, test results and their interpretation, and the exclusion of physical illness) was given the highest priority, and psychiatrists also named the exclusion of physical illness as the minimum they ask for.

Opinions were divided on length. Primary care physicians recognized that “longer is not necessarily better,” but said it was unclear which contents to choose. Psychiatrists were split between those who supported narrowing the letter down to the key points and those who said that more information is better. The effort to keep letters short may have led to insufficient information for some psychiatrists, and the paper proposes prioritizing sufficient information over brevity.

As clues for choosing the contents, participants mentioned writing important information that the patient is unlikely to tell the psychiatrist, and putting detailed information in a separate attachment. They also mentioned asking for permission before writing sensitive content and considering the possibility that the referral letter will be opened.

I think information about the family should be written within the range needed for the purpose of the referral, with attention to the patient’s consent and to confidentiality. Four of the 18 items (the patient’s profile, the family, information from the multidisciplinary team, and the reason for referral) can also serve as a checklist when residents summarize a case.

In the medication history, write the dose and duration as well as the drug name

Primary care physicians said that, in addition to current prescriptions for physical illness and the medication history for psychotropics, they write the changes that followed a prescription, side effects, allergies, and the reasons for changing or stopping a drug. However, some did not know how to write a complicated medication history, and the descriptions ranged from detailed ones that included dose and duration to ones that gave only the drug names.

Psychiatrists wanted a detailed description of doses. To judge the effect of a drug, they check whether the dose and duration used were sufficient. One comment was that some psychotropics require an adequate observation period.

The paper names medication history, along with assessment and diagnosis, as an item on which the two sides’ perceptions differed. I recommend writing the drug name, the dose, the duration of use, the effects and side effects, and the reasons for changing or stopping the drug together. This helps the psychiatrist judge whether the response came after a sufficient dose and duration.

Write your own assessment, even if it is only a symptom name or a provisional diagnosis

The item “assessment and diagnosis” includes the symptom name, the provisional diagnosis, and the diagnosis. Primary care physicians said that they hesitate to write their own assessment and sometimes deliberately leave out a diagnosis. The reasons were the possibility of misdiagnosis and the possibility of biasing the psychiatrist’s view.

Psychiatrists saw it differently. They preferred to have the assessment written and did not expect it to be accurate. They regarded it as their own role to diagnose accurately and objectively, regardless of other physicians’ assessments or diagnoses. The discussion section of the paper also concludes that primary care physicians’ worry that writing an assessment would influence psychiatrists was not borne out.

I appreciate the description of a PCP’s assessment. I am not too concerned about its accuracy because I usually meet the patients with a fresh start. (Psychiatrist)

In a consultation room, a psychiatrist holds a pink sheet of paper in both hands and reads it. A patient sits across a small desk, and a bookshelf stands in the background
Meeting the patient with the referral letter in hand

On how to write, primary care physicians said that, to avoid using psychiatric jargon incorrectly, they describe concrete episodes and quote the patient’s own words. Psychiatrists welcomed this way of writing. They also said, however, that emphasizing only the conspicuous episodes may lead them to misjudge the severity or the purpose of the referral.

On the physical assessment, psychiatrists raised concerns about referral letters that only attached test results and letters that attributed symptoms to psychological problems without giving a basis. What psychiatrists wanted was the line of reasoning: what was suspected, which tests were done, and how the results were interpreted.

Rather than only listing the test results, I would appreciate RLs that only tentatively indicate that the patient needs psychiatric treatment because the test results showed a low possibility of somatic disease. The assertive description of ‘psychological origin’ makes me feel uncomfortable because I doubt it. (Psychiatrist)

I think an assessment should be conveyed together with its basis and course. Write the symptom name or provisional diagnosis along with its basis, the treatments tried, and how they went.

Write what you ask of the psychiatrist and what you will continue yourself

Primary care physicians said that they write whether they plan to keep seeing the patient after the referral. What they wrote in particular was their intention to continue care for existing and newly arising physical illness. Some also write when the psychiatrist can make a reverse referral.

I tried to write ‘After the patient becomes stable, we can follow up with them.’ (…) As much as possible, I try to mention that I did not simply pass the patient on to psychiatrists. (Primary care physician)

Psychiatrists welcomed this description, because a reverse referral is easier to make when they know it is acceptable. One psychiatrist also said that when a referral letter asks for a diagnosis and the current management plan, it is easy to diagnose once, make a plan, and refer the patient back to the primary care physician.

On the reason for referral as well, psychiatrists said it helps when the clinical tasks expected of them are written clearly. Primary care physicians said that they write whether they are asking for hospitalization, for shared care, or for advice on treatment options so that they can continue treatment themselves, so that the intention of the referral is not misunderstood. Some psychiatrists also asked that the reason for the referral and what the patient should consult the psychiatrist about be explained to the patient as well.

I think the line marking how far I manage a patient myself shifts with the illness and the timing. So with each referral, I write what I ask of the psychiatrist and what I will continue myself. One example is the sentence “I will continue to see the patient for physical problems and everyday concerns.” The message comes across more easily when it is specific: whether I will keep managing the physical illness after the referral, and whether I will take over care, including mental health care, once the patient is stable. I also tell the patient the reason for the referral, explaining it in a way that does not leave them feeling abandoned.

As for replies after a referral, primary care physicians said that they do not receive feedback from psychiatrists and that they want to learn from the replies, and a request for a psychiatrist’s reply became one of the 18 items. This study did not examine how often replies arrive or how outcomes are actually reported. I think it is good to state in the referral letter what you want to know, such as the diagnosis and the treatment plan, and for both sides to be able to discuss sharing the patient’s course and dividing roles after the referral.

The 18 items expected in referral letters

From what the participants said, the contents expected to be included in referral letters were organized into 18 items. The list is in Table 2 of the paper1. It is a list for choosing the necessary information according to the case and the purpose of the referral. The main items are the patient’s profile; physical information to facilitate psychiatrists’ practice; family composition, relationship, and dynamics; assessment and diagnosis; description of symptoms by describing a concrete episode or quoting the patient’s words; reason for referral; intention to involve the patient after referral; and medication history.

Psychiatrists asked primary care physicians to avoid jargon and abbreviations unique to a particular specialty. On urgency, it was pointed out that it sometimes has to be conveyed through a channel other than the referral letter.

Attitudes toward writing referral letters were organized into seven themes. In the paper’s wording, they are as follows.

  • Based on relational continuity with PCPs and patients
  • PCP’s initiatives within the scope of their medical practice
  • Making adjustment and considerations to ensure a brief but clear description of the referral purpose and intention
  • Consideration of referral destination’s characteristics and their potential relational impacts
  • Detailed descriptions over the incorrect use of psychiatric jargon
  • Asking for a patient’s permission to include sensitive content
  • Considering the possibility of a referral letter being opened by the patient or their family

On the sharing of clinical information and on AI, the following is my own view. If clinical information can be shared between medical institutions, the burden of copying out prescription histories and test results may decrease. Even then, I think the referral letter will keep the role of conveying together the purpose of this referral, the points to consult about, and the division of roles after the referral. When drafting with AI, use the 18 items for instructions and checking; the physician adds what is not in the records, such as the patient’s state before onset and the current assessment and plan, and checks the draft against the original information.

Limitations of the study

The paper lists two limitations. First, the primary care physicians who participated were limited to family physician specialists. The average primary care physician in Japan may hold different views, such as considering themselves unable to assess or treat mental health problems. Second, although the ideal contents of a referral letter may vary with the referral process, the study did not take such differences into account. As future work, the paper names building a consensus among primary care physicians and psychiatrists on what should be included in referral letters. Differences arising from how clinical information is shared, and writing referral letters with AI, were outside the scope of this study.

Disclosure: I took part in this study as a co-author of the paper. I also serve as chair of the Mental Health Committee of the Japan Primary Care Association.

References

  1. Kajikawa N, Yoshimoto H, Kimura N, So R, Koda M, Yokoya S. A qualitative study on the contents of referral letters from primary care physicians to psychiatrists. BMC Primary Care. 2026. https://doi.org/10.1186/s12875-026-03612-y
English translation of the Japanese article.